Upload
others
View
29
Download
0
Embed Size (px)
Citation preview
Pixel Vertex Detectors
N. WermesBonn University
SLAC SUMMER INSTITUTE 2006
1SSI, 07/20/2006
Outline
1. IntroductionFrom gas-filled chambers to pixel vertex detectors
2. Hybrid Pixel Detectors for the LHCThe Signal and the Noise in Pixel Detector
3. Making a Pixel DetectorFrom sensor to module-ladder
4. Pixel R&D for Future Colliders (addendum)New developments for the ILC
2SSI, 07/20/2006
Outline
1. IntroductionFrom gas-filled chambers to pixel vertex detectors
2. Hybrid Pixel Detectors for the LHCThe Signal and the Noise in Pixel Detector
3. Making a Pixel DetectorFrom sensor to module-ladder
4. Pixel R&D for Future Colliders (addendum)New developments for the ILC
3SSI, 07/20/2006
Introductionimportant advances in tracking through …
multi-wire proportional chambers (1968)and drift chambers (>1972)
electronic recording of tracks σ = mm - 100µm, 0.05 channels / cm2
vertex drift chambers (~1981)vertexing, life times of long lived particlesσ ~ 50µm, 0.1 channels / cm2
silicon micro strip detectors (1983)precision vertexingσ < 10µm, 100 channels / cm2
CharpakSauli
Jaros, Foster, …
Hyams, Weilhammer, Klanner, Lutz
4SSI, 07/20/2006
Introductionimportant advances in tracking through …
multi-wire proportional chambers (1968)and drift chambers (~1975)
electronic recording of tracks σ = mm - 100µm, 0.05 channels / cm2
vertex drift chambers (~1981)vertexing, life times of long lived particlesσ ~ 50µm, 0.1 channels / cm2
silicon micro strip detectors (1983)precision vertexingσ < 10µm, 100 channels / cm2
pixel detectors (since ~1993) tracking and vertexing in LHC environmentσ ~ 10µm, 5000 channels / cm2
5SSI, 07/20/2006
Tracking in pp collisions at 14 TeV (LHC)
~1200 tracks every 25 ns or ~ 1011 per second
⇒ high radiation dose
1015 neq / cm2 / 10 yrs @ LHC
or
600 kGy (60 Mrad)through the ionisation ofmips in 250 µm bulk silicon
pixel detector5000 ch/cm2
LHC ≅ 106 x LEP in track rate !
6SSI, 07/20/2006
Detection tasks of pixel detectors
1. Pattern Recognition and Tracking • precision tracking points in 3D track seeding• 1 pixel layer 3-4 strip layers (x,y & u,v for ambiguities)
2. Vertexing (primary and secondary vertex) 1)
• impact parameter resolution ~10µm (rϕ), ~70µm (z)• secondary vertex resolution ~50µm (rϕ), ~70 µm (z)• primary vertex resolution ~11µm (rϕ), ~45µm (z)• (life) time resolution ~70 fs• (vertex counting luminosity measurement)
3. Momentum measurement 1)
1)values for ATLAS(inner detector)
7SSI, 07/20/2006
10 cm1 cm
Vertexing at LEP …
secondary vertex
primary vertex
Si micro vertex detector~ 10µm resolution
layer 1
layer 2
Z ττ, τ 3π
8SSI, 07/20/2006
Impact parameter resolution (simplified)
beam pipe
9SSI, 07/20/2006
Impact parameter resolution (simplified)
small !small !
small x/X0
beam pipe
10SSI, 07/20/2006
… compare … vertexing at LHC
3D hit informationmandatory
pixels
~ 1200 tracks/BX
high track density in particular in jets
pp ttH (m=120 GeV) H → bbtt → W(lνl)b W(qq)b
11SSI, 07/20/2006
Expected resolutions (ATLAS)
12SSI, 07/20/2006
Expected resolutions (ATLAS)
b-tagging efficiency
light
qua
rk re
ject
ion
13SSI, 07/20/2006
Inner Detector
Tracking Detectors: ATLAS
14SSI, 07/20/2006
points σ (Rφ) (μm) σ (Rz) (μm)pixel 3 12 60SCT 4 17 580
-TRT 36 170
Pixel Detector(3 layers, 3 disks)
Silicon Pixel Detector ~ 1.8 m2
Silicon Strip Detector ~ 60 m2
Transition Radiation Tracker ~ 300 m2eq
50x400 µm2 cells80 x 106 pixels
Tracking Detectors: ATLAS
15SSI, 07/20/2006
Silicon Pixel Detector ~ 1 m2
Silicon Strip Detector ~ 200 m2
Pixel Detector(3 layers, 2 disks)
100x150 µm2 cells33 x 106 pixels
Tracking Detectors: CMS
16SSI, 07/20/2006
Inner Tracking Detectors: ALICE
Silicon Pixel Detector ~ 0.2 m2
Silicon Drift Detector ~ 1.3 m2
Silicon Strip Detector ~ 4.9 m2
Pixel Detector(2 layers, no disks)
50x450 µm2 cells10 x 106 pixels
Rout=43.6 cm
2 strips 2 drifts 2 pixels
+ TPC
17SSI, 07/20/2006
Outline of the Lecture
1. IntroductionFrom gas-filled chambers to pixel vertex detectors
2. Hybrid Pixel Detectors for the LHCThe Signal and the Noise in Pixel Detector
3. Making a Pixel DetectorFrom sensor to module-ladder
4. Pixel R&D for Future CollidersNew developments for the ILC
18SSI, 07/20/2006
Hybrid Pixel Detectors for the LHC
Hybrid Pixel Detector (state of the art)
• amplification by adedicated R/O chip
• 1-1 cell correspondence
19SSI, 07/20/2006
thin (~µm), highly doped p+ (~1019 cm-3) layer on lightly doped n- (~1012 cm-3) substrate
The pn junction as a semiconductor particle detector
neutralitycondition
Maxwell
Space charge region(depleted of mobile carriers)
Electric field
Potential
reverse biased junction
20SSI, 07/20/2006
E
xfull depletion
overdepletion
V ext=V dep
V ext=V dep+ V
The pn junction as a semiconductor particle detectorwith applied external bias voltage
capacitance
full depletion
d
Emax = 2Vext /d
depletion zone grows from the junction into the lower doped bulk
21SSI, 07/20/2006
The Signal and the Noise in pixel detectors
created charge carriers (e/h)move in depletion regionby
drift
and
diffusion
typically 8-10 µm in 300 µm Si
in Si bulk fully depleted• wi = 3.61 eV per e/h• a high energy particle
~ 80 e/h per µm• all charge is collected• ~ 20 000 e/h per 250 µm= 3 fC
• radiatione.g. 10 keV X-ray: 3000 e/h≈ 0.5 fC
note: photo effect ~Z4-5
Si CdTe, CZT, HgI2, …
E
22SSI, 07/20/2006
The Signal and the Noise in pixel detectors
• pixel pattern• typical cells: 100 x 150 µm2
50 x 400 µm2
• charge diffusion σ ~ 8-10 µm• charge spread over 2-4 pixels
note: photo effect ~Z4-5
Si CdTe, CZT, HgI2, …
in Si bulk fully depleted• wi = 3.61 eV per e/h• a high energy particle
~ 80 e/h per µm• all charge collected• ~ 20 000 e/h per 250 µm= 3 fC
• radiatione.g. 10 keV X-ray: 3000 e/h≈ 0.5 fC
23SSI, 07/20/2006
Delta electrons
δ
kinematics: 1-1 relation between emission angleand kin. energy
slow ones emittedat right angles
in 1/β2 part of BBFhighly ionizing
for experimentalists
δ- electrons are “always”emitted at 900 and arehighly ionizing
leadair
Landau distribution
24SSI, 07/20/2006
Delta electrons
single hit clusterδ-electron with perpendicular emission
effect of δ -electrons
100 keV δ-electron occurs in 300 µm Si with 6% probabilityand has “range” of 60 µm
DEPFET pixels (25 µm x 25 µm)
20000 e
250 µm
δ electron
1500 e
typ. 60 µm4500 e
25SSI, 07/20/2006
Signal generation in an electrode configuration
how does a moving chargecouple to an electrode ?• respect Gauss’ law and find
Ramo theorem
weighting field
induction (weighting) potentialdetermines how charge movement couples to a specific electrode
V=1
V=0
26SSI, 07/20/2006
Signal generation in a 2-electrode configuration
d E
V=1
V=0
with
and
27SSI, 07/20/2006
Signal generation in a pixel detector (1-dim)
ΦW for a strip/pixel geometry
y
x
V=1 V=0V=0
V=0
“small pixel effect” !
curre
nt (A
)
time (s)
28SSI, 07/20/2006
Charge collection in a magnetic field
αL
Lorentz angle
perpendicular track
sensor angled w.r.t. track
29SSI, 07/20/2006
Effective incidence angle = tilt angle + Lorentz angle
Lorentz angle measurement (ATLAS)
Deg
Mea
n clus
. size
0
0.5
1
1.5
2
2.5
3
3.5
-20 -15 -10 -5 0 5 10 15 20
Measurement method: number of pixel hits is minimum when incidence angle is equal to the Lorentz angle
Lorentz angle
00
Electric field (V/cm)
Dri
ft v
eloc
ity (c
m/s
)
As bias voltage is increased to cope with irradiation, the Lorentz angle decreases:
Lorentz angle @2T, 150V = -100
Lorentz angle @2T, 600V = -50
Pixel modules tilt in ATLAS = +200
Pixe
l Clu
ster
size
angle
I. Gorelov et al. NIM A481:204-221,2002
30SSI, 07/20/2006
Noise in a pixel detector
three physical noise sources:
number fluctuations of quanta 1. shot noise and 2. 1/f noisevelocity fluctuations of quanta 3. thermal noise
where do they appear in a typical pixel detector readout chain ?
pixel sensor
outin
preamp
white spectrum in frequency f
31SSI, 07/20/2006
three physical noise sources:
number fluctuations of quanta 1. shot noise and 2. 1/f noisevelocity fluctuations of quanta 3. thermal noise
where do they appear in a typical pixel detector readout chain ?
pixel sensor
outin
preamp
thermal transistor channel (white) noise
transistor 1/f noise
Noise in a pixel detector
32SSI, 07/20/2006
three physical noise sources:
number fluctuations of quanta 1. shot noise and 2. 1/f noisevelocity fluctuations of quanta 3. thermal noise
where do they appear in a typical pixel detector readout chain ?
pixel sensor
outin
preamp
thermal transistor channel noise
transistor 1/f noise
Noise in a pixel detector
33SSI, 07/20/2006
equivalent noise charge
referenceRossi, Fischer, Rohe, WermesPixel DetectorsSpringer 2006
Noise in a pixel detector
charge sensitive preamplifier only
34SSI, 07/20/2006
Noise in a pixel detector
… with an additional filter amplifier (shaper) being the band width limiter
opt
shot
~200 fFfor pixels
τ
35SSI, 07/20/2006
typical figures for an LHC pixel detector
Noise = 150 e- initially200 e- after 10 years @ LHC
Signal = 20000 e- total charge in 250 µm Si13000 e- including charge sharing6000 – 8000 e- after 10 yrs @ LHC
S/N > 30
Signal/Noise in a pixel detector
36SSI, 07/20/2006
Outline of the Lecture
1. IntroductionFrom gas-filled chambers to pixel vertex detectors
2. Hybrid Pixel Detectors for the LHCThe Signal and the Noise in Pixel Detector
3. Making a Pixel DetectorFrom sensor to module-ladder
4. Pixel R&D for Future CollidersNew developments for the ILC
37SSI, 07/20/2006
target: 10 years LHC ≅ 1015 neq/cm2 (600 kGy)
Main issue for ATLAS and CMS: Irradiation
• Si sensors: depletion voltage and leakage currents rise• FE chips: threshold shifts & parasitic transistors occur• glue: becomes hard and brittle• mechanics: material performance degrades• cooling: larger capacity is needed to cool more power
intensive irradiation and test beam program over yearsincluding dedicated high intensity beams with LHC likerates and timing structure
Note: Plans for Super – LHC (~2015): SLHC = LHC x 10
38SSI, 07/20/2006
50 μm
ATLAS Modul, Foto:IZM, Berlin
Sensors• n+ in n (oxygenated Si) • wafer size (Ø 10 cm)• ~200-250 µm thick
Electronics - Chip• chip size limited by yield ~1-2.5 cm2
• wafer size (Ø 20 cm)Hybridization
• PbSn or Indium bumps (wafer scale)• IC wafers thinned after bumping to ~180 µm• ‚flip-chip‘ to mate the parts• ~3000 bumps/chip, ~50000 bumps/module
Hybrid Pixel Assembly
FE-I3 wafer map, 82% yieldATLAS pixel BARE module
PbSn
50 µmIZM,Berlin
ATLAS pixelsensor wafer
ATLAS FE-I3 Wafer
CMS Pixel Modul
CMS module
(with Flex Hybrid and Controller Chip TBM)
50 µm
Indium (lift-off)
In
AMS Rome
Inreflow
PSI
sensor
chip
39SSI, 07/20/2006
flex-hybrid
pigtail
MCCsensor
FE-chip
out to opto interface
bare
assembled
Hybrid Pixel Assembly
40SSI, 07/20/2006
Plasma activation
EvaporatedIndium
Wet Lift off process
Photolithography
Wafer CleaningProcess parameters:
• Resist Thickness: 15 μm
• Pre-bake: 30min @ 80 °C
• Deposition rate: 0.5 μm/min
• Dep. Pressure: 9 x 10 - 7 Torr
• Temp. during Dep. < 50 °C
Indium bumping process
Flip-Chip
41SSI, 07/20/2006
Solder bumping process
Flip-Chip
42SSI, 07/20/2006
Pixel Sensors in the LHC radiation environment
particle interactions with lattice nuclei
NIELnon-ionizingenergy loss(not reversible)normalized to1 MeV neutron damage
recoiling Si-atom can cause further defectsdefect clusters (10nm x 200nm)
1. generation/recombination levels in band gapincrease of leakage current
2. change of space charge in depleted regionchange of effective doping concentration
3. trapping centers created trapping of signal charge
vacancy
doublevacancy
impurityinterstitialimpurity
atominterstitial
Frenkeldefect
hadron
43SSI, 07/20/2006
Pixel Sensors in the LHC radiation environment
Change of Depletion Voltage Vdep (Neff)…. with particle fluence:
• “Type inversion”: Neff changes from positive to negative (Space Charge Sign Inversion)
10-1 100 101 102 103
Φeq [ 1012 cm-2 ]
1
510
50100
5001000
5000
Ude
p [V
] (d
= 3
00μm
)
10-1
100
101
102
103
| Nef
f | [
1011
cm
-3 ]
≈ 600 V
1014cm-2
type inversion
n-type "p-type"
[M.Moll: Data: R. Wunstorf, PhD thesis 1992, Uni Hamburg]
fluence (NIEL) > 1015 neq/cm2
total dose > 500 kGy
particle interactions with lattice nuclei
NIELnon-ionizingenergy loss(not reversible)normalized to1 MeV neutron damage
recoiling Si-atom can cause further defectsdefect clusters (10nm x 200nm)
1. generation/recombination levels in band gapincrease of leakage current
2. change of space charge in depleted regionchange of effective doping concentration
3. trapping centers created trapping of signal charge
44SSI, 07/20/2006
0 0.5 1 1.5 2 2.5 3 3.5Φeq [1014cm-2]
1
2
3
4
5
6
7
|Nef
f| [1
012 c
m-3
]
100
200
300
400
Vde
p [V
] (3
00μm
)
neutronsneutrons
neutronsneutronspionspionsprotonsprotons
pionspionsprotonsprotons
oxygen rich FZ
standard FZstandard FZ
radiation tolerant to 1015 neq / cm2 (600 kGy)
typeinversion
Pixel Sensors in the LHC radiation environment
solution: oxygenated FZ silicon
protonspionsneutrons
necessary voltage for full depletion
RD50, G. Lindström et al.NIM-A 465 (2001) 60-69
45SSI, 07/20/2006
Pixel Sensors in the LHC radiation environment
21
21
3
Change of Depletion Voltage Vdep (Neff)…. with particle fluence:
• “Type inversion”: Neff changes from positive to negative (Space Charge Sign Inversion)
10-1 100 101 102 103
Φeq [ 1012 cm-2 ]
1
510
50100
5001000
5000
Ude
p [V
] (d
= 3
00μm
)
10-1
100
101
102
103
| Nef
f | [
1011
cm
-3 ]
≈ 600 V
1014cm-2
type inversion
n-type "p-type"
[M.Moll: Data: R. Wunstorf, PhD thesis 1992, Uni Hamburg]
fluence (NIEL) > 1015 neq/cm2
total dose > 500 kGy
p+ in n “normal” n+ in n
before type inversion
after type inversion
after heavy radiation damageL. Andricek et al, NIM-A 409 (1998) 184-193
undepletedundepleted
46SSI, 07/20/2006
maximum E-field
Pixel Sensors: isolation of pixel implants
p-stop
highest E-fields after irradiation
p-spray
E-fields decrease with irradiation
moderated p-spray
optimum configurationfor overall voltage stability
maximum E-field maximum E-field
R. Richter et al, NIM-A 377 (1996) 412
47SSI, 07/20/2006
Biasing of Pixel Sensors
(up to 600 V)
48SSI, 07/20/2006
pixel implant
contact pad
Biasing of Pixel Sensors
49SSI, 07/20/2006
punch through biasing
below equal aboveVpunch through Vpunch through Vpunch through
V
pixelpixelpixel
Biasing of Pixel Sensors
50SSI, 07/20/2006
for testing
ATLAS
Bias grid
Punch through dot
Bump contact
P-spray
low fields afterirradiation
n-pixelsisolation by
moderated p-spray
see paperby R. Richter
~ homogeneous charge collection after 10 years LHC
after irradiation to 1015 neq/cm2
Biasing of Pixel Sensors
SSI, 07/20/2006
variable angleof incidence
Scintillator+PMs
Barrel module
August 2005 CMS beam test
Measuring the effective depletion depth after irradiation
52SSI, 07/20/2006
V.Chiochia, M. Swartz et al. arxiv.org/abs/physics/0409049, IEEE NSS 04 in Rome
• Shallow angle method
CMS
Measuring the effective depletion depth after irradiation
Cluster charge, all clusters
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
5 10 15 20 25 30 35 40 45 50
200 V400 V600 V
ke1/
N*d
N/d
Q 18000e
cluster charge
ATLAS
after 600 kGy
45minsi
gnal
hei
gth
(arb
. uni
ts)
cluster size
53SSI, 07/20/2006
Trapping after 10 yrs @ LHCUse tilted tracks again …
For non-irradiated sensors,thecollected charge is uniform alongthe depth.
The charge yield yield as afunction of the depth can betranslated, via the driftvelocity, in a carrier lifetime:
τe = 4.1 ± 0.3 ± 0.5 ns
m)μTrack depth (0 50 100 150 200 250
Cha
rge
(arb
itrar
y un
its)
0
0.2
0.4
0.6
0.8
1
Charge Vs Depth
Bonn3, irradiated, 600 V (run1857)
Bonn9, not irradiated, 150 V (run 1333)
Charge Vs Depth
Irradiated
not-irradiated
A. Andreazza et al
mean CCE after 10 yrs LHC ~ 80%(with LHC type annealing scenario)
ATLAS
54SSI, 07/20/2006
Pixel Frontend Chip
ATLAS FE-I3• 0,25 µm CMOS technology• pixel cell size: 50 x 400 µm2
• 18 columns x 160 rows = 2880 cells• parallel processing in all cells
- amplification- zero suppression 11
mm
7.4 mm
~700 transistors per pixel cell~ 3.5 M transistors total
L. Blanquart, P. Fischer et al., NIM-A 456 (2001) 217-231Active voltage drop
compensationPreamplifier 5-bit DAC – global
thresholdRAM write/read
controlAuto-tune
SEU-tolerant RAM cells for pixel configuration
bits7-bit tune-DAC for local threshold adjustmentBonding-pad
50 μm
400 μm
55SSI, 07/20/2006
Functions in the cell (binary readout + „poor man‘s“ analog)
- Integration of signal charge by charge sensitive amplifier- Pulse shaping by feedback circuit with constant current feed back- Hit detection by comparator - ~5 bit analog information via „time over threshold“- storage of address and time stamps in RAM at the periphery
Feedback
(6+1)-bitlocal threshold
DAC
Calibrationcharge
injection
StrobeCalibration voltage
Select
Hit
Global threshold
Bump bond contact
AddressROM
FallingedgeRAM
LeadingedgeRAM
Global time stamp(40 MHz gray counter)
Prioritylogic
Bus to column controller
Hit data & Arbitration logic
tr tf
ToT
L. Blanquart et al., NIM-A 456 (2001) 217-231
56SSI, 07/20/2006
Requirements on the electronics performance
• small noise hit rate low noise and small threshold dispersion• σnoise ⊕ σthreshold < ~ 600 e- @ a threshold of 3000 e-
• time stamp < 20 ns after BX for all signal heights
Distribution of pixel cell thresholds
no tuning
with tuning
57SSI, 07/20/2006
Important / in-time threshold & efficiency
zero time walk linefor large signals
set threshold e.g. to 4000 ± 200 e-
(lowest possible ~1500 e-)
In-time threshold (Δt<20ns): 5200 ± 200 e-
⇒ overdrive = 1200 ± 200 e-
(achieved in irradiated assemblies)
in-time efficiency ~99% wanted and achieved !
timewalk
58SSI, 07/20/2006
11mm
7.4 mm
ATLAS FE-Chip• 0,25 µm CMOS technology• pixel cell size: 50 x 400 µm2
• 18 columns x 160 rows = 2880 cells• parallel processing in all cells
- amplification- zero suppression
• end of column logic- storage of hit information during
trigger latency (2.5 µs)- hit selection upon L1 trigger
Pixel Frontend Chip
L. Blanquart, P. Fischer et al., NIM-A 456 (2001) 217-231
59SSI, 07/20/2006
40 MHz Gray coded clocktransmitted to all cells
Pixel cells generate hit information (address and time stamp) which are stored at the end of column
hits are removed if no trigger conicidence occurs
Hit information agreeing with L1 trigger time are read out
I. Peric
ATLAS Pixel Chip: binary hit information with additional information on signal hight via ToT measurement (~4-5 bit)
Analogue circuitsDigital readout circuitsRegisters used to store configuration bitsTime informationTrigger
ALTAS FE-chip readout architecture (animated)
60SSI, 07/20/2006
CMS pixel-chip (analog readout)
functional blocks similar to ATLAS Pixel-Chip FE-I3additional storage of analog pulse height (sample/hold)analog output signal amplitude + row/column address coded in analog levels
CMS Pixel-Chip PSI46V2
H.C. Kastli et al., e-print physics/0511166
61SSI, 07/20/2006
pixel address decoded into binary numbers for DAQ
• Overlay of 4160 pixel readouts (analog coded address levels)
5 clock cycles encode 13 bits of pixel address information.
analog pixel pulse height
1 pixel hit
address
puls
e he
ight
CMS pixel-chip (analog readout)
ultra black
H.C. Kastli et al., e-print physics/0511166
62SSI, 07/20/2006
Radiation damage to the FE-electronics … and cure
Effects: generation of positive charges in the SiO2 and defects in Si - SiO2 interface
1. Theshold shifts of transistors
DSM CMOS technologies with small structure sizes (≤ 0,35 µm) and thin gate oxides (dox < 10 nm) holes tunnel out
2. Leakage currents under the field oxideLayout of annular transistors with annular gate-electrodes + guard-rings
p-Substrat
n+ n+
Drain Source
GateGate-Oxid
Feld-Oxid
Leckstrom
Source
Gate
Drainp-Substrat
n+ n+
Drain Source
Gate Gate-Oxid
+ +
++
particle/radiation
++
++--
--
--
++
63SSI, 07/20/2006
radiation induced bit errors (“single event upsets“ SEU)
large amounts of charge on circuit nodes- by nuclear reactions, high track densities -can cause “bit-flip“
2 examples of error resistant logic cells
enlarge storage capacitances in SRAM cells: Qcrit = Vthreshold · C
storage cells with redundancy (DICE SRAM cell)
Q-
1 0
SEU tolerant DICE SRAM cell
0 1
standard SRAM cell
Radiation damage to the FE-electronics … and cure
64SSI, 07/20/2006
before irradiation
<ENC> = 152e
σthr = 40e
noise thr
Irradiated Modules after 1 MGy (20 years @ LHC)
after 1 MGy
20yrsLHC <ENC> = 182e
σthr = 127e
ATLAS
noise thr
ATLAS Pixel Modul (without Flex)
A. Andreazza NIM-A 513:103-106,2003
65SSI, 07/20/2006
Spatial resolution in irradiated assemblies
before irradiation
(100 incl.)
after 60 Mrad
(100 incl.)
ATLAS
66SSI, 07/20/2006
efficiencyno hitout of time plateaumasked
97.8%1.5%0.7%9.7 ns~10-4
25ns
large in-time plateau for efficency margin
efficiency vs time, standard
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50 60 70ns
effic
ienc
y
eff. 99.9%no hit 0.1%out of time 0%plateau 14ns
25ns
In-time track efficiency in irradiated assemblies
before irradiation after 600 kGy
ATLAS
67SSI, 07/20/2006
Main issue for ALICE: minimal material
Primary charged particles in a central event
In central HI collisions up to 8000 charged particles/ |η| are expected.
radiation levels only ~ 5 kGy, 6x1012 neq / cm2
operation at room temperature possible !
~80 hits/cm2
68SSI, 07/20/2006
Main Issue for ALICE: minimal material
very light weight Carbon Fibresupport structure (200µm,0.1 X0)sensor 200µmIC 150µmcooling (C4F10) @ RT 0.3% X0(PHYNOX tubes, wall 40µm)
total X0 per layer ~ 0.9%
ALICE
(ATLAS, CMS > 2%)
69SSI, 07/20/2006
Hybrid Pixels / Ladders and Disks
- minimal X0 “C-C” structures- cooling (pumped C3F8 : boil. point = -250) - T < -60 C to limit damage from irradiation- power dissipation: ~100W/stave
(ATLAS) ~15kW/detectorALICE
ATLAS
70SSI, 07/20/2006
Hybrid Pixel Detectors @ LHC
complex signal processing in cells • zero suppression • temporary storage of hits during L1 latency
radiation hardness to 1015 neq/cm2
spatial resolution ~ 10–15 µm
… but also
relatively large material budget: ~ 2% X0 per layer (1% X0 @ ALICE)• cooling, services
complex and laborious module production • bump-bonding / Flip-Chip expensive• many production steps
71SSI, 07/20/2006
Conclusionshybrid pixel detectors are the “state of the art” of pixel vertex detectors
spin-offs into imaging applications are abound• X-ray pixel detectors (MPEC, MEDIPIX, CiX)• X-ray astronomy (DEPFET, CdTe pixels)• time resolved autoradiography• … many more
next challenges are around the corner • Super-LHC
radiation fluences up to 1016 neq/cm2 new sensor types“light weight” less power, new cooling, new mechanicsdata band width 40 MHz >GHz
• Monolithic pixel detectors for ILC(semi)-monolithic pixel detectors: MAPS, DEPFETnew technologies: SOI pixels, a-Si:H pixels
Join in ! There is enough to do !
72SSI, 07/20/2006
Further Reading Rossi, Fischer, Rohe, Wermes, “Pixel Detectors: From Fundamentals to Applications”,Springer Berlin-Heidelberg-New York, 2006, (ISBN 3-540-283324)
G. Lutz, “Semiconductor Radiation Detectors”, Springer Berlin-Heidelberg-New York, 1999.
E. Heijne, “Semiconductor Micropattern Pixel Detectors: A Review of the Beginnings”, NIM A465(2001) 1-26
N. Wermes, “Pixel Detectors for Tracking and theirs Spin-off in Imaging Applications”Nucl.Instrum.Meth.A541:150-165,2005, e-Print Archive: physics/0410282and “Pixel detectors”, in LECC2005 Heidelberg 2005, Electronics for LHC and future experiments e-print Archive: physics/0512037
ATLAS Pixel Detector, Technical Design Report, CERN/LHCC/98-13 (1998)CMS Tracker Technical Design Report, CERN/LHCC/98-6 (1998)ALICE Inner Tracker System, Technical Design Report, CERN/LHCC/99-12 (1999)
R. Horisberger, “Readout Architectures for Pixel Detectors”, NIM A465 (2001) 148-152L. Blanquart et al., “Pixel Readout Electronics for LHC and Biomedical Applications”, NIM A439(2000) 403-412
73SSI, 07/20/2006
Outline
1. IntroductionFrom gas-filled chambers to pixel vertex detectors
2. Hybrid Pixel Detectors for the LHCThe Signal and the Noise in Pixel Detector
3. Making a Pixel DetectorFrom sensor to module-ladder
4. Pixel R&D for Future Colliders (addendum)New developments for the ILC
74SSI, 07/20/2006
R&DCCDMAPSDEPFETSOI …
Requirements
• Thin (< 50 µm, 0.1% X0) monolithic• > 500 Mpix with small cells (< 25x25 µm2 ) • Fast (50 MHz/line, 25 kHz/frame ≈2Mpix)• Low power (few Watts for full detector)• Radiation tolerance < 4 kGy = 1/25 of LHC• No trigger
Pixel Detectors for a Linear Collider
Time structure and rates
• 80 hits / mm2 / bunch train @ r=1.5 cm
b/c tagging
75SSI, 07/20/2006
Principle of (semi-) monolithic pixel detectors
generation and integration of signal in same substrate• pn-diode QSignal
• collection diode (transistor gate) USignal = QSignal / Cg
or ISignal = gm · QSignal / Cg
• row wise selection of pixels • column wise readout• switch in cell (select/reset)
MAPS (CMOS active pixels)• same CMOS substrate (low resistivity) forsteering/readout electronics and Q - collection
DEPFET• amplifying transistor on fully depleted bulk (high resistivity), separate steering and R/O chips
pixel cell
select line
read out line
row
sele
ctio
nan
d c
lear
column readout
pixel matrix
76SSI, 07/20/2006
Principle of (semi-) monolithic pixel detectors
generation and integration of signal in same substrate• pn-diode QSignal
• collection diode (transistor gate) USignal = QSignal / Cg
or ISignal = gm · QSignal / Cg
• row wise selection of pixels • column wise readout• switch in cell (select/reset)
MAPS (CMOS active pixels)• same CMOS substrate (low resistivity) forsteering/readout electronics and Q - collection
DEPFET• amplifying transistor on fully depleted bulk (high resistivity), separate steering and R/O chips
DEPFET
77SSI, 07/20/2006
Conclusionshybrid pixel detectors are the “state of the art” of pixel vertex detectors
spin-offs into imaging applications are abound• X-ray pixel detectors (MPEC, MEDIPIX, CiX)• X-ray astronomy (DEPFET, CdTe pixels)• time resolved autoradiography• … many more
next challenges are around the corner • Super-LHC
radiation fluences up to 1016 neq/cm2 new sensor types“light weight” less power, new cooling, new mechanicsdata band width 40 MHz >GHz
• Monolithic pixel detectors for ILC(semi)-monolithic pixel detectors: MAPS, DEPFETnew technologies: SOI pixels, a-Si:H pixels
Join in ! There is enough to do !
78SSI, 07/20/2006
Further Reading Rossi, Fischer, Rohe, Wermes, “Pixel Detectors: From Fundamentals to Applications”,Springer Berlin-Heidelberg-New York, 2006, (ISBN 3-540-283324)
G. Lutz, “Semiconductor Radiation Detectors”, Springer Berlin-Heidelberg-New York, 1999.
E. Heijne, “Semiconductor Micropattern Pixel Detectors: A Review of the Beginnings”, NIM A465(2001) 1-26
N. Wermes, “Pixel Detectors for Tracking and theirs Spin-off in Imaging Applications”Nucl.Instrum.Meth.A541:150-165,2005, e-Print Archive: physics/0410282and “Pixel detectors”, in LECC2005 Heidelberg 2005, Electronics for LHC and future experiments e-print Archive: physics/0512037
ATLAS Pixel Detector, Technical Design Report, CERN/LHCC/98-13 (1998)CMS Tracker Technical Design Report, CERN/LHCC/98-6 (1998)ALICE Inner Tracker System, Technical Design Report, CERN/LHCC/99-12 (1999)
R. Horisberger, “Readout Architectures for Pixel Detectors”, NIM A465 (2001) 148-152L. Blanquart et al., “Pixel Readout Electronics for LHC and Biomedical Applications”, NIM A439(2000) 403-412
79SSI, 07/20/2006
AddendumPixel Detectors
for ILC
80SSI, 07/20/2006
CMOS Active Pixels
at n-well
in epi-layer
• charge coll. in several µm thin epi-layer by thermal diffusion to n-well/epi junction• p-wells and substrate highly doped charges kept between reflection boundaries• signals processed by standard CMOS circuitry integrated on sensor • only nMOS in active area (due to n-well/epi collection diode)• Q-collection time ~100 ns (due to diffusion)• incomplete Q-collection and small signals (< 1000 e) => challenge for IC design• small pixel sizes (< 20x20 µm2): a must and a virtue !
Diericks, Meynants, Scheffer (1997)
L. Jungermann
Meynants, Diericks, Scheffer, SPIE 3410:68-76 (1998)R. Turchetta, NIM-A 458:677-689 (2001)
81SSI, 07/20/2006
CMOS active pixels / R/O & performance
select
„standard“ 3 transistor R/O scheme→ upgraded to include amplification, current memory (15 transitors, MIMOSA-7)
resetsource follower stage
row selection → column R/0
• small signal (< 1000e)=> low noise needed
• detectors sizes up to19.4 x 17.4 mm2 (1Mpix)
• smallest pitch: 17 µm
• spatial resolution < 2µm
82SSI, 07/20/2006
from D. Contarato
83SSI, 07/20/2006
internal Gate
Potential distribution:
[TeSCA-Simulation]
DEPFET pixels: high ohmic bulk
p+
p+ n+
n
n+
totally depletedn--substrate
internal gate
rear contact
source top gate drain bulk potential via axistop-gate / rear contact
V
potential minimumfor electrons
p-channelp+
--
- -+
+
++
-- --1 μm
50 µm- 300 μm
(MOS)FET-Transistor integrated in every pixel (first amplification)Local potential minimum (for e- ) under transistor channelElectrons are collected in „internal gate“ and modulate the transistor-currentSignal charge removed via clear contactoutput is a current
J. Kemmer und G. Lutz;, NIM A253 (1987) 365
84SSI, 07/20/2006
internal Gate
Potential distribution:
[TeSCA-Simulation]
Monolithic Pixels / DEPFET pixels
p+
p+ n+
rear contact
drain bulksource
p
sym
met
rya x
is
n+
ninternal gate
top gate clear
n -
n+p+~1µm
50 µ
m- -- ---
0V
+15V
0V
CLEAR
(MOS)FET-Transistor integrated in every pixel (first amplification
small CD (fF) => very low noise (< 2e- achieved in spectroscoopy devices, ~100e- @ ILC)large signal => thin detectors (50 µm) S/N = 40-80 @ ILClow power => ~ few watts for entire detector (5 layers) save cooling (X0)
detector sizes: 64 x 128 pixels, ~25x25 µm2 cells
85SSI, 07/20/2006
Current Readout chip(read currents, store and subtract pedestals)
- Read cells of a row& store their currents
- Clear internal gatesof this row completely
- Read again (pedestalcurrents) and subtractby adding currents
Operation of a DEPFET Matrix• 1 active row
DEPFETs are ON
R/O CLEAR R/O
• all other rows OFFstill active for signals
low power !
86SSI, 07/20/2006
ILC Detector Concept
50µm
300µm
DEPFET
matrix
Switcher Chips
CUROs
thinned sensor diode
50µm 300µm
minimal cooling (X0) required~ 5W for 5 layer VTX detector
layer 1
87SSI, 07/20/2006
Backup Slides
88SSI, 07/20/2006
FE-chip wafer yields (0.25 µm CMOS)
ALICE (SPD-RO)CMS (ROC)ATLAS (FE-I3)
82% 51%
11 x 7.4 mm2
180µm thick
yieldsbefore thinning
7.9 x 9.8 mm2
200µm thick13.5 x 15.8 mm2
150µm thick
> 80%
89SSI, 07/20/2006
E-fields in “p-stop” and “p-spray” in comparison
90SSI, 07/20/2006
~90% produced @ IZM & AMS - ~ 2x20 modules/week - rework fraction : 10% - 15% - rework efficiency:
solder ~100%, indium ~80%- module reject fraction:
solder ~ 1%, indium ~14%
Hybrid Pixels / BARE module yield (ATLAS)
0.1% total need (3 layers): 1744 + sparestotal order @ bump vendors: ~2500delivered (20.1.2006): ~2200fully assembled (today): ~2000
91SSI, 07/20/2006
Hybrid Pixels / module quality yield
Ranking levels:b-layer, layer 1, layer 2
ranking based on:- inefficient pixel- sensor quality- noise performance- threshold tuning- rebonding- BareModule rework
overall ranking
failed: 120; 10%
layer2: 256; 21%
layer1: 180; 15%
b-layer: 668; 54%
b-layerlayer1layer2failed
total: 1225
layer2
layer1
b-layer
93SSI, 07/20/2006
Power consumption
• R/O Chip: 2mW / channel 200 W (whole vtx-d)
Number of R/O channels @ TESLA:L1 : 520x2x8 = 8320L2-5: (880x2)x(8+12+16+20) = 98560
All: 106880 channels
• Sensor: PDEPFET = 5V x 100 µA = 500µW 50 W
• Steering: 0.94mW /channelDC, 3.13mW / channel @ 50MHz
L1 : 2x3.13 + (3998x0.94) mW= 34WL2-5: [2*3.13 + (13538x0.94)] x (8+12+16+20) mW = 713W
All: 747 W
Total : 997W , 1/199 duty cycle : 5W